Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A characteristic analytics system comprising a computer system, the computer system comprising one or more processing units and a memory coupled to at least one of the one or more processing units, the memory comprising instructions for: a) receiving, in electronic form, a communication that provides information, wherein the communication is selected by a user and is received in the form of a data construct that provides the information in a first format associated with a source of the communication; b) formatting the data construct from the first format to a second format, wherein the second format comprises a text object that contains the information, thereby forming the text object; c) applying a first subset of classification models in a plurality of classification models to the text object, wherein each respective classification model in the first subset of classification models parses a portion of the text object in accordance with a corresponding plurality of heuristic instructions associated with the respective classification model into a respective classification in the form of a corresponding plurality of classification text strings, and wherein the corresponding plurality of classification text strings of the respective classification collectively contains a portion, less than all, of the information, thereby forming a plurality of classifications, each classification in the plurality of classifications corresponding to application of a classification model in the first subset of classification models to the text object; d) evaluating, for each respective classification model in a second subset of classification models, each classification in the plurality of classifications, by applying each respective classification model in the second subset of classification models to each corresponding plurality of classification text strings of each respective classification in the plurality of classifications, in which each respective classification model in the second subset of classification models applies a corresponding plurality of heuristic instructions selected by the respective classification model based on a determination of the source of the communication, thereby forming a corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; and e) providing, in electronic form, a characteristic of the communication in the form of a result of the evaluating d), wherein the characteristic is an amalgamation of the corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; wherein the first subset of classification models and the second subset of classification models collectively comprise three unique classification models in the plurality of classification models, and wherein the first subset of classification models and the second subset of classification models are disjoint subsets of classification models in the plurality of classification models.
A characteristic analytics system processes electronic communications to extract and evaluate information. The system receives a user-selected communication in a source-specific format, converting it into a standardized text object. A first set of classification models analyzes the text object, each model parsing portions of the text using heuristic instructions to generate partial classifications as text strings. These classifications collectively contain only part of the original information. A second set of classification models then evaluates each classification by applying heuristic instructions selected based on the communication's source. The evaluations from the second set are combined to produce a final characteristic of the communication. The system uses three distinct classification models, with the first and second sets being non-overlapping subsets. This approach enables structured analysis of diverse communication formats by leveraging specialized models for parsing and evaluation, ensuring accurate extraction of key characteristics.
2. The system of claim 1 , wherein the first format of the data construct comprises audio data, and wherein the formatting b) further comprises transcribing the audio data of the communication, thereby forming the text object.
This invention relates to a system for processing communication data in different formats, particularly focusing on audio data. The system converts audio data from a communication into a text object through transcription, enabling further processing or analysis. The core functionality involves receiving communication data in a first format, such as audio, and applying formatting operations to transform it into a second format. For audio data, this includes transcribing the audio into text, creating a text object that can be stored, analyzed, or used in other applications. The system may also handle other data formats, but the key innovation lies in the transcription step for audio, ensuring accurate conversion while preserving the original communication's context. This approach improves accessibility, searchability, and integration with text-based systems, addressing challenges in handling unstructured audio data in digital workflows. The system may include additional components for data storage, retrieval, or further processing, but the transcription of audio into text is a central feature. This method enhances efficiency in applications like voice-to-text services, meeting transcriptions, or automated documentation.
3. The system of claim 2 , wherein the transcribing is conducted using a speech recognition classification model in the plurality of classification models.
This invention relates to a system for processing audio data, specifically focusing on transcribing speech using machine learning models. The system addresses the challenge of accurately converting spoken language into written text, particularly in scenarios where multiple classification models are employed to handle different aspects of the audio data. The system includes a plurality of classification models, each specialized for distinct tasks such as speech recognition, speaker identification, or language detection. One of these models is a speech recognition classification model, which is specifically tasked with transcribing the audio input into text. The system dynamically selects and applies the appropriate models based on the characteristics of the audio data, ensuring accurate and context-aware transcription. The speech recognition model processes the audio input, converting spoken words into written text while accounting for variations in speech patterns, accents, and background noise. This approach enhances the reliability and efficiency of automated transcription systems, making them suitable for applications such as real-time captioning, voice assistants, and automated meeting notes. The system's modular design allows for easy integration of additional models to improve performance across different audio environments.
4. The system of claim 1 , wherein the information provided in the first format comprises an ideogram.
The invention relates to a system for processing and displaying information in a user interface, particularly for enhancing user interaction with complex data. The system addresses the challenge of presenting information in a way that is both visually intuitive and easily interpretable, especially when dealing with large or intricate datasets. The system converts information from a first format into a second format for display, where the first format includes an ideogram—a graphical symbol that represents an idea or concept. This ideogram is transformed into a more detailed or structured representation in the second format, allowing users to interact with the data in a more meaningful way. The system may also include a user interface that facilitates navigation and manipulation of the transformed data, ensuring that users can efficiently extract insights. The ideogram serves as a compact, recognizable representation that simplifies the initial presentation of information, while the second format provides deeper context or additional details. This approach is particularly useful in applications where clarity and ease of understanding are critical, such as data visualization, user interfaces for software applications, or educational tools. The system ensures that users can quickly grasp the essence of the information before delving into more detailed aspects.
5. The system of claim 4 , wherein the ideogram comprises an image emoticon, a text emoticon, or a combination thereof.
The invention relates to a system for generating and displaying ideograms, such as emoticons, in digital communication. The system addresses the challenge of conveying emotions or reactions in text-based communication by providing a dynamic and customizable way to represent feelings through visual or textual symbols. The ideograms can be generated based on user input, context, or predefined templates, and may include images, text-based emoticons, or a combination of both. The system ensures that the ideograms are easily recognizable and adaptable to different communication platforms, enhancing user engagement and emotional expression in digital interactions. The ideograms may be integrated into messaging applications, social media, or other digital interfaces to improve the richness of communication. The system also allows for real-time adjustments to the ideograms, ensuring they remain relevant and contextually appropriate. By offering a flexible and intuitive way to express emotions, the system enhances the overall user experience in digital communication environments.
6. The system of claim 1 , wherein the plurality of classification models comprises a decision tree classification model and wherein a corresponding plurality of heuristic instructions associated with the decision tree classification model comprises a plurality of pre-pruning instructions, a plurality of post-pruning instructions, or a combination thereof.
This invention relates to a machine learning system that improves the performance of classification models, particularly decision tree models, by applying heuristic instructions to optimize their structure and accuracy. The system addresses the challenge of balancing model complexity and generalization in decision trees, where overfitting or underfitting can degrade performance. The invention includes a plurality of classification models, with at least one being a decision tree model, and a corresponding set of heuristic instructions designed to refine the tree's construction. These instructions include pre-pruning techniques, which limit the tree's growth during training to prevent overfitting, and post-pruning techniques, which simplify the tree after training to improve generalization. The system dynamically applies these heuristics to enhance the decision tree's predictive accuracy and efficiency. By integrating these optimization steps, the invention ensures that the decision tree model achieves a better trade-off between complexity and performance, making it more reliable for real-world applications. The approach is particularly useful in scenarios where interpretability and computational efficiency are critical, such as in automated decision-making systems.
7. The system of claim 6 , wherein the plurality of heuristic instructions associated with the decision tree classification model comprise a plurality of information gain heuristic instructions.
A system for improving decision-making in machine learning models, particularly for classification tasks, addresses the challenge of optimizing model performance by selecting the most informative features. The system includes a decision tree classification model that processes input data to generate predictions. The model is enhanced by a plurality of heuristic instructions that guide feature selection and branching decisions within the tree structure. These heuristic instructions are specifically designed to maximize information gain, a metric that quantifies the reduction in uncertainty or entropy when a feature is used to split the data. By prioritizing features with the highest information gain, the system ensures that the decision tree makes more accurate and efficient splits, leading to improved classification performance. The heuristic instructions may include rules for evaluating feature importance, determining optimal split points, and handling missing or noisy data. The system is particularly useful in applications where interpretability and efficiency are critical, such as healthcare diagnostics, financial risk assessment, and fraud detection. The use of information gain heuristics ensures that the model focuses on the most relevant features, reducing computational overhead and enhancing decision reliability.
8. The system of claim 1 , wherein the characteristic of the communication comprises an emotion of the communication, a sentiment of the communication, or a combination thereof.
This invention relates to a system for analyzing communication characteristics, particularly focusing on detecting and processing emotional and sentimental aspects of communications. The system is designed to evaluate interactions between users, such as in messaging platforms, social media, or customer service environments, to identify emotional states or sentiments expressed in the communication content. By analyzing these characteristics, the system can enhance user engagement, improve response accuracy, or tailor interactions based on detected emotional cues. The system may use natural language processing, machine learning, or other computational techniques to extract and interpret emotional or sentimental indicators from text, voice, or other communication modalities. The analysis can involve classifying communications into categories such as positive, negative, or neutral sentiment, or identifying specific emotions like happiness, frustration, or urgency. The system may then apply these insights to automate responses, prioritize communications, or trigger alerts for further action. The goal is to provide more context-aware and empathetic communication handling, improving user satisfaction and operational efficiency in various applications.
9. The system of claim 1 , wherein the characteristic of the communication comprises an emotion of the communication in the form of happiness, sadness, fear, disgust, anger, surprise, pride, shame, embarrassment, and excitement.
This invention relates to a communication analysis system that identifies and processes emotional characteristics within communications. The system detects and categorizes emotions expressed in communications, such as happiness, sadness, fear, disgust, anger, surprise, pride, shame, embarrassment, and excitement. The system analyzes communication data, such as text, speech, or other forms of interaction, to extract emotional content. It then processes this emotional data to provide insights, such as sentiment analysis, emotional tone detection, or behavioral assessment. The system may be used in applications like customer service, mental health monitoring, or social media analysis, where understanding emotional context is critical. By identifying and quantifying emotions, the system enhances communication understanding, enabling more effective responses or interventions. The system may integrate with other technologies, such as natural language processing or machine learning, to improve accuracy and adaptability in recognizing diverse emotional expressions.
10. The system of claim 1 , wherein the characteristic of the communication comprises an emotion of the communication in the form of admiration, adoration, aesthetic appreciation, amusement, anxiety, awe, awkwardness, boredom, calmness, confusion, craving, disgust, empathetic pain, entrancement, envy, excitement, fear, horror, interest, joy, nostalgia, romance, sadness, satisfaction, sexual desire, sympathy, and triumph.
The system relates to analyzing and processing communication data to identify and categorize emotional characteristics within the communication. The problem addressed is the need to accurately detect and interpret the emotional content of communications, such as text, speech, or other forms of interaction, to improve user engagement, personalization, or system responses. The system includes a communication analysis module that processes input communication data to extract and classify emotional attributes. These attributes are mapped to a predefined set of emotions, including admiration, adoration, aesthetic appreciation, amusement, anxiety, awe, awkwardness, boredom, calmness, confusion, craving, disgust, empathetic pain, entrancement, envy, excitement, fear, horror, interest, joy, nostalgia, romance, sadness, satisfaction, sexual desire, sympathy, and triumph. The system may use machine learning, natural language processing, or other analytical techniques to determine the emotional tone or sentiment of the communication. The system may further include a response generation module that tailors outputs based on the detected emotions, such as generating empathetic responses, adjusting system behavior, or providing personalized recommendations. The emotional analysis can be applied in various domains, including customer service, mental health support, social media interactions, or virtual assistants, to enhance user experience and interaction quality. The system may also track emotional trends over time to identify patterns or shifts in user sentiment.
11. The system of claim 1 , wherein the characteristic of the communication comprises a combination of one or more emotions of the communication.
A system analyzes communication data to extract and process emotional characteristics. The system identifies multiple emotions present in a communication, such as text, speech, or other forms of interaction, and combines these emotions to determine an overall emotional profile. This profile can be used to assess the emotional state of the communicator, improve communication quality, or enhance user experience in applications like customer service, mental health monitoring, or social media analysis. The system may employ natural language processing, speech recognition, or other analytical techniques to detect emotions from linguistic patterns, tone, or other indicators. By evaluating the combined emotional content, the system provides insights into complex emotional states that may not be apparent from individual emotions alone. This approach helps in applications requiring nuanced emotional understanding, such as conflict resolution, personalized recommendations, or adaptive user interfaces. The system may also integrate with other communication analysis tools to refine emotional detection accuracy or contextualize findings.
12. The system of claim 1 , wherein the characteristic of the communication comprises a sentiment of the communication that is a positive sentiment, a neutral sentiment, a negative sentiment, or a combination thereof.
This invention relates to a communication analysis system that evaluates characteristics of communications, particularly focusing on sentiment analysis. The system processes communications to determine whether they exhibit a positive, neutral, or negative sentiment, or a combination of these sentiments. By analyzing sentiment, the system can assess the emotional tone or attitude conveyed in the communication, enabling applications such as customer feedback analysis, social media monitoring, or automated response systems. The system may integrate with other components, such as data collection modules or machine learning models, to enhance accuracy and provide actionable insights. The sentiment analysis can be applied to various communication types, including text, speech, or multimedia, and may involve natural language processing techniques to classify sentiment accurately. The system helps organizations understand user emotions, improve engagement, and make data-driven decisions based on sentiment trends.
13. The system of claim 1 , wherein the data construct is derived from the communication, and wherein the source of the communication comprises a social media communication feed, an email communication, a telephonic communication, or a technical document.
This invention relates to a system for processing and analyzing communications from various sources to derive structured data constructs. The system is designed to extract meaningful information from unstructured or semi-structured communication data, enabling improved data organization, retrieval, and analysis. The communication sources include social media feeds, emails, telephonic conversations, and technical documents. The system processes these inputs to generate a standardized data construct that captures key elements of the communication, such as sender, recipient, content, timestamps, and context. This structured output facilitates automated workflows, data integration, and decision-making processes. The system may also include components for filtering, categorizing, or enriching the derived data constructs based on predefined rules or machine learning models. By transforming diverse communication formats into a unified structure, the system enhances interoperability and usability across different applications, such as customer support, business intelligence, or compliance monitoring. The invention addresses challenges in handling fragmented communication data by providing a scalable and adaptable solution for extracting actionable insights.
14. The system of claim 1 , wherein the second format is in accordance with a standardized format.
A system for data processing involves converting data between different formats to ensure compatibility and interoperability. The system includes a data conversion module that transforms data from a first format to a second format, where the second format adheres to a standardized format. This standardization ensures that the converted data can be seamlessly integrated into systems or applications that require compliance with specific industry or technical standards. The data conversion module may include preprocessing steps to prepare the data for conversion, such as cleaning, validation, or normalization, to ensure accuracy and consistency. The system may also include a validation module to verify that the converted data meets the requirements of the standardized format, reducing errors and improving reliability. By enforcing standardized formats, the system enhances data exchange efficiency, reduces compatibility issues, and supports regulatory compliance in industries where standardized data formats are mandatory. The system can be applied in various domains, including finance, healthcare, and telecommunications, where data must adhere to predefined standards for interoperability and security.
15. The system of claim 1 , wherein the plurality of classification models comprises a decision tree classification model, a neural network classification model, a support vector machine classification model, a Naïve Bayes classification model, a pattern-matching classification model, a syntactic based classification model, or a combination thereof.
The invention relates to a system for classifying data using multiple machine learning models. The system addresses the challenge of accurately categorizing data by leveraging diverse classification techniques to improve robustness and adaptability. The core system includes a plurality of classification models, each designed to process input data and generate classification outputs. These models may include a decision tree classifier, which uses a tree-like structure to make decisions based on feature values; a neural network classifier, which employs interconnected layers of nodes to learn complex patterns; a support vector machine classifier, which separates data points using hyperplanes; a Naïve Bayes classifier, which applies probabilistic methods based on Bayes' theorem; a pattern-matching classifier, which identifies predefined patterns in the data; and a syntactic-based classifier, which analyzes structural or grammatical features. The system may combine these models to enhance accuracy, reduce bias, or handle different types of data. By integrating multiple approaches, the system improves classification performance across various applications, such as text analysis, image recognition, or anomaly detection. The flexibility in model selection allows customization based on specific use cases, ensuring adaptability to different data types and requirements.
16. The system according to claim 1 , wherein the plurality of classification models comprises a neural network classification model in the form of an inter-pattern distance based classification model.
A system for classification tasks utilizes multiple classification models, including a neural network-based inter-pattern distance classification model. The system is designed to improve accuracy and robustness in classification by leveraging different model architectures. The neural network classification model operates by computing distances between input patterns and known reference patterns, enabling classification based on similarity metrics. This approach enhances the system's ability to handle complex data distributions and improve decision-making in applications such as image recognition, natural language processing, or anomaly detection. The inter-pattern distance method allows the model to generalize better across diverse datasets by focusing on relational features rather than rigid feature extraction. The system integrates this neural network model with other classification models to provide a comprehensive and adaptive classification framework. The combination of models ensures that the system can adapt to various data types and classification challenges, improving overall performance and reliability. This approach is particularly useful in scenarios where traditional classification methods may fail due to data variability or noise. The system's modular design allows for easy integration of additional models or updates to existing ones, ensuring scalability and future-proofing.
17. The system of claim 1 , wherein: a first corresponding plurality of classification text strings of a first classification model in the first subset of classification models is associated with a semantic classification, and a second corresponding plurality of classification text strings of a second classification model in the first subset of classification models is associated with a syntax classification.
The invention relates to a system for classifying text data using multiple classification models, each specialized in different types of classifications. The system addresses the challenge of accurately categorizing text by leveraging distinct classification approaches, such as semantic and syntactic analysis, to improve overall classification performance. The system includes a first subset of classification models, where each model processes text data to generate classification results. Within this subset, a first classification model is configured to associate a plurality of classification text strings with semantic classification, focusing on the meaning and context of the text. A second classification model in the same subset associates a different plurality of classification text strings with syntactic classification, which examines the structure and grammatical rules of the text. By combining these models, the system enhances the accuracy and robustness of text classification tasks, accommodating both meaning-based and structure-based analysis. The system may also include additional subsets of classification models, each with specialized functions, to further refine classification outcomes. The integration of diverse classification approaches allows the system to handle complex text data more effectively, ensuring comprehensive and precise categorization. This method is particularly useful in applications requiring high accuracy, such as natural language processing, document management, and automated content analysis.
18. A method for analyzing characteristics of a data construct, the method comprising: a) receiving, in electronic form, a communication that provides information, wherein the communication is selected by a user and is received in the form of a data construct that provides the information in a first format associated with a source of the communication; b) formatting the data construct from the first format to a second format, wherein the second format comprises a text object that contains the information, thereby forming the text object: c) applying a first subset of classification models in a plurality of classification models to the text object, wherein each respective classification model in the first subset of classification models parses a portion of the text object in accordance with a corresponding plurality of heuristic instructions associated with the respective classification model into a respective classification in the form of a corresponding plurality of classification text strings, and wherein the corresponding plurality of classification text strings of the respective classification collectively contains a portion, less than all, of the information, thereby forming a plurality of classifications, each classification in the plurality of classifications corresponding to application of a classification model in the first subset of classification models to the text object; d) evaluating, for each respective classification model in a second subset of classification models, each classification in the plurality of classifications, by applying each respective classification model in the second subset of classification models to each corresponding plurality of classification text strings of each respective classification in the plurality of classifications, in which each respective classification model in the second subset of classification models applies a corresponding plurality of heuristic instructions selected by the respective classification model based on a determination of the source of the communication, thereby forming a corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; and e) providing, in electronic form, a characteristic of the communication in the form of a result of the evaluating d), wherein the characteristic is an amalgamation of the corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; wherein the first subset of classification models and the second subset of classification models collectively comprise three unique classification models in the plurality of classification models, and wherein the first subset of classification models and the second subset of classification models are disjoint subsets of classification models in the plurality of classification models.
The invention relates to analyzing characteristics of digital communications by processing data constructs in different formats. The problem addressed is the need to extract meaningful information from communications that may be in various formats, such as emails, documents, or other structured or unstructured data, and to classify and evaluate that information using multiple models tailored to the communication's source. The method involves receiving a user-selected communication in a first format, such as an email or document, and converting it into a standardized text object. A first set of classification models processes this text object, each model parsing a portion of the text according to its own heuristic rules to generate partial classifications. These classifications are then evaluated by a second set of classification models, which apply additional heuristic rules based on the communication's source. The results from these evaluations are combined to produce a final characteristic of the communication. The system uses three distinct classification models, with the first and second sets being non-overlapping. This approach ensures that different aspects of the communication are analyzed independently before being merged into a comprehensive result. The method is designed to improve the accuracy and relevance of information extraction by leveraging multiple specialized models.
19. A non-transitory computer readable storage medium stored on a computing device, the computing device comprising, one or more processors and memory storing one or more programs for execution by the one or more processors, wherein the one or more programs singularly or collectively comprise instructions for running an application on the computing device that executes a method comprising: a) receiving, in electronic form, a communication that provides information, wherein the communication is selected by a user and is received in the form of a data construct that provides the information in a first format associated with a source of the communication; b) formatting the data construct from the first format to a second format, wherein the second format comprises a text object that contains the information, thereby forming the text object: c) applying a first subset of classification models in a plurality of classification models to the text object, wherein each respective classification model in the first subset of classification models parses a portion of the text object in accordance with a corresponding plurality of heuristic instructions associated with the respective classification model into a respective classification in the form of a corresponding plurality of classification text strings, and wherein the corresponding plurality of classification text strings of the respective classification collectively contains a portion, less than all, of the information, thereby forming a plurality of classifications, each classification in the plurality of classifications corresponding to application of a classification model in the first subset of classification models to the text object; d) evaluating, for each respective classification model in a second subset of classification models, each classification in the plurality of classifications, by applying each respective classification model in the second subset of classification models to each corresponding plurality of classification text strings of each respective classification in the plurality of classifications, in which each respective classification model in the second subset of classification models applies a corresponding plurality of heuristic instructions selected by the respective classification model based on a determination of the source of the communication, thereby forming a corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; and e) providing, in electronic form, a characteristic of the communication in the form of a result of the evaluating d), wherein the characteristic is an amalgamation of the corresponding evaluation for each respective classification model in the second subset of classification models for each classification in the plurality of classifications; wherein the first subset of classification models and the second subset of classification models collectively comprise three unique classification models in the plurality of classification models, and wherein the first subset of classification models and the second subset of classification models are disjoint subsets of classification models in the plurality of classification models.
This invention relates to a system for processing and classifying electronic communications using machine learning models. The system addresses the challenge of extracting meaningful characteristics from unstructured or semi-structured data received from various sources. The method involves receiving a communication in a first format, converting it into a standardized text object, and applying multiple classification models to analyze the content. A first set of models parses the text into partial classifications, each containing a subset of the original information. A second set of models evaluates these classifications, selecting heuristic instructions based on the communication's source. The results from both model sets are combined to produce a final characteristic of the communication. The system ensures comprehensive analysis by using three distinct models, with no overlap between the first and second model subsets. This approach enables accurate interpretation of diverse communication formats while maintaining flexibility in processing rules based on source-specific heuristics. The invention is implemented on a computing device with processors and memory, executing programs to perform the described operations.
Unknown
November 3, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.